View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Effects of Automation on the Structure and Functioning of Banking

Remarks of George W. Mitchell
Member, Board of Governors of the Federal Reserve System
at the
Annual Meeting of the
American Economic Association
New York City
December 28, 1965

Effects of Automation on the Structure and Functioning of Banking

The commercial banking system in the United States owes its
existence to its ability to render for all sectors of the economy a
unique and pervasive service, namely, the movement of money.


transfer of funds between debtor and creditor, buyer and seller,
citizen and government, employer and employee--in fact from anyone with
a balance in a bank to any identifiable payee is ever-ready, safe and
Banks consummate money settlements through an elaborate clearing
system for transferring deposit balances from one account to another in
the same bank or in different banks.— ^
ordinary check.

The principal tool used is the

The mechanical and institutional channels by which

60 million checks daily find their way physically, and with appropriate
accounting entries, from any one to any other among 70 million accounts
is vastly more complicated than might be thought.

It achieves a high

degree of ultimate accuracy considering its predominant dependence on
manual processing.

The service is far from instantaneous, however,

being no faster than present-day methods of transporting checks in
considerable bulk about a city or throughout the country.
In the aggregate, the annual cost of operating the present
check and settlement system for the entire country can be estimated at
about 3.3 billion dollars.

Some of this cost is borne by the Federal

Reserve System, some is paid in the form of service charges assessed
on the account of the payor, and some is "clipped" from the face of the
check instrument itself and charged the payee when nonpar banks are

However, the bulk of the total cost is absorbed by the

commercial banks themselves to attract balances that can be invested
in interest-earning assets.



Given the cost and time pressures inherent in the present check
settlement system, it is not surprising that the initial bank efforts to
exploit the infant prodigy of electronic data processing have been aimed
mainly at lowering the per-item handling cost and speeding up the per­
formance of old-style check collection.

This is constructive, so far as

it goes, but it falls far short of the potential transformation of
banking services rendered possible by automation.

Laying aside all the

undoubted strains and pains of the transition, it is practical to
envision the advancement of the state of the art to a point that will
permit--and perhaps almost force--radical change in banking structure
and functions.

This state will be reached within the discernible future,

probably much sooner than most of us expect.
By that time, I expect, check usage as we know it will have
largely disappeared, and the intricate process of settlement and deposit
accounting will be carried on concurrently at and between 250 or so

computer centers located throughout the country.—

A modified giro

system will be used, in which the payor will initiate the settlement
process, but will do so by communicating, not with the payee, but with
his bank--notifying it directly whom to pay, how much, and when.
Most of this information will be received at the bank in
machine language; if not, it will be converted to that form, and the
bank's computer will process the bookkeeping entries internally for
amounts drawn on it.

If one computer handles the accounts for several

banks the operation is s t i l l ^ l m ^ t entirely an internal one.
is to an account

If payment



routed into that bankfs equipment.

Bank positions will also be adjusted

frequently throughout the day by debits and credits to member bank
accounts with the Federal Reserve System.

The computers will transmit

printed-out confirmations to the payor and advices to the payee at
appropriate intervals.

The print-outs could be transmitted by mail

or telephone wire, at the option of the customer.

In the case of larger

customers, the bankfs computer will communicate directly with customerfs
In this system there is no check sorting and re-sorting, no
shipment of checks from bank-to-bank or bank-to-customer, no storage
requirements for checks, no kited checks, no endorsement, no N.S.F.
checks, no float, and a minimum of manual processing.
different problems may later come to light.

Of course,

The machine must work;

and the bank must make sure it is being instructed by the owner of the

There is no reason, however, to fear that any such potential

difficulties are beyond the technological capacities and probably the
cost horizons now in view.
Furthermore, it seems logical and practical that at least
some of the customer accounting antecedent and subsequent to settlement
could be most economically done in a coordinated package with the settle­
ment accounting.

Every sales transaction, for example, by specifying a

settlement date, might immediately be put into the bank’ computer
where it could accomplish immediate settlement or subsequent reminder
and settlement.

Similarly, a bank could handle payrolls and agree to

bill and process many types of contractual payments for insurance, rent,


and mortgage payments.


In short, by virtue of its central position in

the payments process the bank is also able to perform ancillary and
antecedent accounting and billing operations more economically than
anyone else.
For willing business customers, the bank's service could
include a large part of the accounting, analysis, and financing of
receivables and even extend to provision of much current cash flow
accounting, a basis of analysis that has become of increasing importance
in both business and financial planning.
Tied into the possibility of, if not a prerequisite to,
expanded service for most depositors is the introduction of a depositor
combination cash/credit card.

This device could be used for immediate

payment, partially replacing the use of coin and currency, or it could
be used for the processing of convenience credit or the scheduling and
liquidation of instalment or revolving credit.—

All of these ancillary operations enhance profitable business
prospects of a computerized settlement system outlined above.


they and similar extensions of service have an important bearing upon
the alacrity and enthusiasm with which banks will convert to or adopt
EDP systems.
If the foregoing projections are realistic, they seem to
promise, in the aggregate, a substantially more efficient settlement

And they imply additional profit opportunities for banks

that can combine settlement with receivables accounting, payroll
accounting, credit card operation, and a consumer credit system
for depositors.


Today, no one really knows how much cost reduction, private
and social, a fully computerized system might achieve*

Some of those

with the earliest and most extensive experience in partial EDP applica­
tions are taking a hard, if not skeptical, look at "hardware" costs of
a full-scale operation.

One of the difficulties of bringing the

relevant evidence together is the problem of totaling up actual private
costs in our present settlement system which could be eliminated or
reduced in the "checkless-cashless" society.

Another is the allocation

of the new system's public and private costs among various public and
private beneficiaries.

If public costs presently involved in the dis­

tribution of currency and processing of checks, for example, are sub­
stantially reduced how can the entrepreneur banker who brings this about
collect a quid pro quo?
The question then of whether and how commercial banks will
regard EDP as an opportunity for profitable service is not easily

The banking system is not distinguished for its innovative

achievements--despite evidence of improvement in recent years.


adaptability to change is hampered by regulatory constraints on structure
and function.

Entry, branching, and merger are closely regulated as are

prices paid for deposits and many of the conditions under which credit
can be extended.

Thus, there is a tradition of conservatism in manage­

ment reinforced by competitive

sheltering and regulatory constraints

that act as inhibitions to innovative steps with any evident structural
or functional consequences.

-6This opens the way for nonbank enterprises to become well
established in the EDP record-keeping applications antecedent or
related to settlement before banks even enter the field.

If they do,

their customers are likely to be relatively indifferent to tardily
offered adjuncts to a banking settlement system.
The deferred entry of banks into the consumer credit business
is a case in point.

Today, credit for consumers is relatively independent

of the banking system though it is also available there, and generally
for less.

However, even today, seldom does a bank's consumer credit

system exploit the natural advantages of a continuing depositor relation­
ship in the way that vendors, for example, have exploited the continuing
patronage of their customers.

Moreover, consumer credit is a natural

extension of other banking operations.

A bank's individual depositors

use credit extended by its retail firm depositors with the proceeds of
trade-credit bank lines.

Vendors have established practical standards

of creditworthiness for their customers and profitably priced the credit

At the most, very few banks have offered an aggressively

competitive alternative to vendor credit by carrying the credit financing
from producer or wholesaler through to the ultimate consumer.
As an industry, banks have moved into consumer credit far more
slowly than vendors; they have not been innovative and, by and large,
have made almost no use of their key position in settlement accounting
to provide services more broadly and economically than anyone else.
Past experience suggests, therefore, that the banking system
may well be reluctant or inhibited from exploiting the opportunities



inherent in automation, particularly if more aggressively-minded EDP
machine sellers or users can carve out large sectors of potential
service in which, through lower costs and innovative flexibility, they
can establish customer loyalties.
But if skepticism borne of experience leads me to expect
opportunities to be missed in this area, logic compels me to insist
that such a fate is not inevitable.

A clear enough vision of future

possibilities exists in some banks today.

Hammered home hard enough

by both intra-industry communication and private advice, it might
serve to erode much of the inertia and inhibition now forestalling a
full-fledged revolution in the settlements mechanism.
If this should happen--and assuming public policies are
accommodative--not only banking services but also banking structure
could be literally transformed.
almost a certainty.

Profound structural changes seem

Automation can and will burst the locational

constraints that are implicit in Federal conformity to the provisions
of 50 State banking laws pertaining to branching.

Not only will

metropolitan area-wide banking operations become commonplace everywhere
at the option of bank managements but remote control banking State-wide,
and even across State lines, will also be feasible, limited by little
more than the telephone toll costs of servicing more distant customers.
To be sure, banks have detoured branching restrictions for
some time, as large banks in various parts of the country have solicited
and made loans all over the United States--or over the world, for that
matter--and have accepted deposits by mail or wire from customers


wherever located.


But this sort of substitute for branching is a

reality only for large accounts.

The dynamic change that will come

into being is that computerizing the demand depositor-bank relationship
will make it practicable and in all probability economically profitable
for banks spectacularly to extend their present service areas for small
and medium-sized accounts.

This they can do, by using the U.S. mails

or by hooking their computer onto a local telephone in any community
they wish to serve.
The same features of automation that will enable banks to
achieve many of the advantages of a far-flung branching system will
also introduce a large element of obsolescence into many existing
branch facilities.

Branches that have been established to achieve

proximity to depositors and are essential from the bank's standpoint
only because they facilitate the sweeping up of loanable funds or
minimize deposit fluctuations by more nearly encompassing the local
payments cycle will become superfluous.
Depositors will have no need to visit their banking office
any more often than they now visit their telephone or electric utility
company office.

They will not be making deposits of checks; rather

their bank will notify them of credits to their accounts.

They will

rarely find it necessary to go to their bank to obtain cash, even
for transactions that are now typically made with cash.
line will be activated automatically.
the equivalent of cash at

The coin and curre

Their credit

Their cash/credit card will be
:et, the cleaners, or a department
r transactions that will



continue to be most conveniently handled in that fashion will be
supplied from commercial establishments that are regularly serviced
by money truck pickup and delivery.
Perhaps there will be a place for "baby branches"--small
field offices which might serve as headquarters for account salesmen
and loan officers, and for performing custodial, certification and
routine financial advisory services, but it is hard to visualize the
typical branch office in existence today as fitting into a computerized
banking institution of the future.
Automation in banking will likely have lesser effects on such
traditional types of bank credit extension as farm lending, mortgage
lending, and loans to large businesses and to financial enterprises.
But it should have a major impact upon consumer credit and trade
credit between firms, particularly of small and medium size.


types of credit involve substantial investigative and bookkeeping
costs relative to interest earned on a typical loan or line of credit.
They also involve more surveillance and more losses, though of a
readily insurable type, thus adding further to overhead costs.
With automation banks can offer a credit system which ties
settlement accounting into quasi^automatic credit extension; this
combination has great operating advantages over other arrangements
available to vendors or independent consumer finance companies.


bank depositor credit card is of superlative convenience for the
purchaser when he can use it anywhere and in doing so express his
preference for cash payment, convenience credit, instalment credit
or any combination of the three.


The bank, in offering this service, can extend credit to
seller or buyer, or both, on the basis of prearranged lines, lines
that have been fixed with access to unparalleled sources of informa­
tion on the customer’ financial activity and responsibility. Moreover,
the computer continuously updates this information and can alert the
bank's credit department on a timely basis to the emergence of credit
abuses by whatever standard the bank may choose to employ.


the convenience of a continuing scrutiny of the customer's cash inflow
and outflow in relation to use of bank credit, and all monitored by a
sentry who reports instantaneously.1
Such a system would not be without losses to be sure, but
they could be controlled by fixing maximum credit lines for various
types of accounts.

And loan limits could serve another purpose--that

of fostering larger demand deposit balances.

If a line of credit,

for example, were some multiple of average daily balance it is quite
likely that most depositors would gladly pay the "commitment fee" for
the convenience and prestige of bank credit.

And still another

advantage so far as banks are concerned is the conventional preference
of many bankers for self-liquidating short-term credit, met in this
instance by the rapid turnover of consumer and sales credit of the
type envisioned.
Particular beneficiaries of this more flexible and better
disciplined credit use should be the many self-employed in the economy,
ranging from part-time salesmen to proprietors.

While their aggregate

credit use is a small factor in bank lending, a significant public
interest is served by uncovering any method of economically and


conveniently making more bank resources available to them.


current operating procedures, the overhead costs associated with such
credit, when added to a regular interest charge, entail effective
interest rates that are prohibitive or appear highly discriminatory.
Automation offers a method of minimizing overhead costs and probably
reducing risk, thereby making bank credit more accessible to a sector
of the economy that has found all sources of credit "high


In their continuous search for loanable resources banks


find automation a far keener tocl than they are accustomed to using.
On the one hand, it will enable them to attract demand deposit customers
with assurance of a simpler, safer, and more convenient means of pay­
ment than has ever been offered.

Not only will the computer reduce the

risk in paying bills, it will also take over the chores in banking-such as a trip to the bank and the standing in line to make a deposit,
the writing of checks and mailing them to creditors, and similar
routine tasks.

Moreover, it will give the depositor "instant bookkeeping"

as he will be able to find out as often as he likes the exact status of
his account, and with that knowledge give the bark instruction as to whom
to pay, and when.,

Thus he is enabled to mar age his money position as

closely as he likes.
From the standpoint of profitable operations, barks that
offer a service making possible the close management of customers1
bank accounts are almost certain to find their dema d deposit totals
wasting away as a manifestation of the automation program.


given the capabilities of a computerized economy, more frequent


settlement periods are likely, if not certain, to come into widespread
use and this development will diminish still further the size of a
comfortable operating balance for the typical depositor.

Just as

weekly wage and salary payments go with a lower operating bank balance
than is needed when payments are monthly, the shorter interval made
possible by automation will call for even smaller cash balances.
Given declining need for demand deposit balance for these
"technological1 reasons, banks have the alternative of establishing
fees to cover at least a portion of the costs incurred for processing
flows through demand deposit accounts or of establishing minimum
balances commensurate to the scope and cost of services rendered.
If they rely heavily on fees, then in order to maintain their aggregate
of loans and investments banks will need to attract time deposits
(or to borrow) in one form or another in competition with other banks,
other financial intermediaries, and the capital markets.

The compensating

balance alternative, on the other hand, if enhanced in appeal by linking
it to a packaged credit line, would enable banks to minimize losses in
demand deposit balances.
Other changes consequent to the automation of money flows-such as operating space and labor requirements of banking institutions-involve formidable housekeeping and management adjustments, but they
are of a different order of concern.
Nor has any mention been made of the possibility of the
settlement system being nationalized, along the lines of European
experience, in the Post Office or the Federal Reserve System.


-13such a step is technically feasible, if not advantageously suited
to a monopolistic operation, our preferences run strongly against
extending Government operations into service areas that can be
satisfactorily performed privately.
If the views and speculations advanced here are at all
persuasive, it will probably be with the assenter's proviso that
l it won't happen in my time.1 To this skepticism I can only reply
that most of the innovations I have alluded to are now in being, or
about to be placed in operation.

Individual banks in all sections

of the country are adopting, piecemeal, elements of a system such
as has been described.

Before very long these experimental operations

will provide a solid foundation for the new banking system of the


14 -

If Commercial banks also have an important function to perform in the
only alternative procedure for settling the economy's accounts, i.e.,
by the use of coin and currency.

They maintain at significant cost

local reservoirs of cash for the use of their businesses and individual
customers who can withdraw amounts and denominations when required and
return to the pool unfit or excess holdings.
27 The number of computer centers given is conjectural but compromises an
optimum operational size and a convenient geographical area.

The very

rapid growth of computer centers recently seems to presage the automation
of all demand deposit accounting in the near future--using checks or
any other settlement media.

Banks will have a choice of their own

equipment, a correspondent's facilities, a cooperative processing
organization or a commercial service bureau.
3/ The cash card involves some exposure to theft or counterfeiting but
various identification devices such as voice "fingerprinting," or other
technological developments now under study appear adequate to control

Security may be more importantly achieved, however, by the

positive identification of all payees inherent in the system.

To be a

payee one must have an account with a bank and have met whatever
identification and responsibility standards are found to be necessary
for the protection of payors.

There are no intermediate signatories

in the system--once the payor has directed his bank to make a given

-15Footnotes (Continued)
payment and the bank is satisfied with its identification of the payor,
then there is no opportunity of intercepting and misappropriating the
"document1 used for this purpose.